Investigating an automated method for the sensitivity analysis of functions

Size: px
Start display at page:

Download "Investigating an automated method for the sensitivity analysis of functions"

Transcription

1 Investigating an automated metod for te sensitivity analysis of functions Sibel EKER Jill SLINGER Delft University of Tecnology 2628 BX, Delft, te Neterlands Gonenc YUCEL ABSTRACT Automated sensitivity analysis approaces in system dynamics focus primarily on model parameters. Altoug table functions are often subjectively approximated, tey do not form te focus of most sensitivity analyses. Recently, a promising approac tat allows automation of sensitivity analysis on functions was proposed by Hearne (2), but te applicability of tis metod to system dynamics table functions as not been studied, yet. In tis study, te metod is applied to a simple system dynamics model. In te ligt of te observations a number of sortcomings are identified and a set of extensions to address tese are proposed and ten tested. Te results of experiments wit te original and te extended metod demonstrate tat te metod can be used easily and efficiently for table functions. Te extensions are sown to be valuable in creating a more compreensive metod, but tey also raise te researc issue of te tradeoff between teir added value and te cost of dealing wit increased complication. Apart from our experimental results, te article also puts fort a set of directions along wic te approac can be improved furter. Despite te issues requiring furter researc, te metod olds promise for routine implementation. KEY WORDS: function sensitivity analysis, automated sensitivity analysis, epidemic model, triangular functions, uncertainty, table functions, system dynamics. INTRODUCTION Te problems modeled using system dynamics are caracterized by uncertainty, arising from a lack of information on te system itself or due to conflicting opinions of te actors involved. Tis real life uncertainty is reflected in model building as uncertainty in te selection of parameter values or in expressions for variable interactions. Despite te difficulties associated wit formulating tem, models ave to be robust/insensitive to suc uncertainties to be considered valuable and useful in solving real life problems. Sensitivity analysis is broadly defined as te study of model responses to model canges (Tank-Nielsen 98), and provides us wit a way to deal wit uncertainty. Elaborately, te importance of sensitivity analysis stems mainly from four factors. Firstly, it enances understanding about bot te structure-beavior relationsip in te

2 model and real world. Secondly, and as te main reason for its common usage, it sows te effects of uncertainties in te model, e.g. estimated parameters, on te conclusions derived from tis model. Tirdly, it determines te parameters wic affect te beavior strongly and so facilitates te devotion of limited resources to estimate tem. Lastly, te parameters wic ave a strong effect on te beavior may be te key points upon wic policies can be built (Sterman 2, 83, Tank-Nielsen 98). However, system dynamics as been criticized in te past owing to te absence of a precise teory or metod to conduct sensitivity analysis (Meadows 98). Following te explication and initial refutation of tis critique by Meadows, various approaces ave been developed. Several autors addressed te importance of sensitivity analysis, and provided basic non-automated approaces to conduct it. In teir early work, Ford, Amlin and Backus (983) empasized te difficulties of sensitivity analysis suc as large number of parameters to cange and large number of state variables wose responses are to be observed; and listed te approaces developed by tat time. Later on, studies are focused on coping wit tese problems. One of te two main approaces was to use optimization to determine parameter values wic cause maximum deviation from te original model beavior (J. Hearne 987, Miller 998). Te second approac wic attracted more attention was te use of statistics to determine combinations of parameter values in multivariate sensitivity analysis, and to interpret te responses of state variables (Kleijnen 995, Clemson, et al. 995). Anoter example of use of statistics was presented by Ford and Flynn (25) were tey screened te model and utilized te correlation coefficients to find te parameters most influential on te beavior. Meanwile, te progress in software tecnology allowed system dynamics and simulation packages to include automated sensitivity analysis tools. Tis reduced te burden of a compreensive sensitivity analysis to some extent, but te abovementioned statistical approaces are still needed. In addition to tose, since te sensitivity is defined as beavioral sensitivity rater tan numerical sensitivity in system dynamics, metods to recognize canges in te beavior patterns ave been developed and incorporated wit sensitivity analysis. Recent examples of suc studies can be seen in (Yucel and Barlas 2) and (Hekimoglu and Barlas 2). One common aspect of tese studies is tat tey deal primarily wit sensitivity of te model to canges in parameter values or to te canges in some model structure features suc as te model boundary or te level of aggregation (Sterman 2, 884). In system dynamics, nonlinear relationsips between two variables are usually specified by a lookup or table function, wic sows ow te dependent variable nonlinearly varies as te independent variable canges (Sterman 2, 552). In te context of model building, table functions are sometimes considered as parameters (Tank-Nielsen 98), but te attention paid to tem even in parameter sensitivity analysis is very limited. However, model robustness can depend on te coices of table functions as well because te sapes of tem are only approximations of real nonlinear relationsips, and due to tat sensitivity analysis is indeed te final step of formulating a table function (Sterman 2, ). Currently, sensitivity analysis activities, particularly automated analyses, focus on parameter sensitivity. Even if a modeler wonders about te outcome of is coices and carries out a manual sensitivity analysis for functions, tis effort will be limited because varying te functions manually, especially in combinations of multiple functions, is cumbersome. Clearly, an automated metod for function sensitivity analysis is required, before te analysis of te effects of uncertainty in functions can join parameter sensitivity analysis as a routinely applied tool. 2

3 Recently, Hearne (2) proposed a novel metod to conduct sensitivity analysis of model functions. Tis metod is based on systematic and parametric perturbation of te grapical functions. Despite being a promising proposal, tis approac as not yet been tested torougly nor as its wider applicability been establised. Te objective of tis study is two-fold. First, tis study aims to explore te applicability of Hearne s recent work on te sensitivity analysis of system dynamics functions. Tis is undertaken by implementing te metod on a simple model wit a single table function, and identifying te concerns and issues requiring attention. Second, extensions to te metod are proposed in order to overcome te identified sortcomings. Te delineated extensions are ten implemented and tested on te same model, and teir usefulness is assessed. Wit tis intention, before diving into te sensitivity analysis metod, te model tat is used in tis study is first introduced in te next section. Tereafter, te tird section deals wit Hearne s metod and its application on te cosen model. In Section 4, te proposed extensions are described and implemented. Te paper ends wit a discussion of te results and a conclusion regarding te promise of te metod. 2. THE SAMPLE MODEL Te model selected for testing te metod for te sensitivity analysis of functions is te basic form of te epidemics model (SIR model) wit an alternative infection rate formulation. In te original model (Sterman 2, 33) were S stands for susceptible population and I is te infected population. Te infection rate (IR) is defined as Te beavior yielded by tis formulation wen te contact rate is 6 people per person per day, te infectivity is.25, te initial number of Infected is and te Total Population is is depicted in Figure. person S I R 2. W2 W3 W4 W5 W6 Non-commercial use only! Figure : Te beavior of original SIR model over time In te alternative formulation te infection rate defined as 3

4 and f depicts te non-linear relationsip between te ratio of I to te total population and te infection rate. Tis relationsip is linear in te original formulation. To ensure similarity to te original formulation, te initial sape of tis table function is specified as depicted in Figure 2, wile te model beavior obtained wit tis function is depicted in Figure 3. Figure 2: Te table function f used in te alternative SIR model person S I R 2. W2 W3 W4 W5 W6 Non-commercial use only! Figure 3: Te beavior of alternative SIR model over time Two assumptions embedded in tis table function will be adopted in tis study, too. First, te function starts at (,), wic means tat wen tere are no infected people in te community, ten tere is no infection. In oter words, tis disease is transmitted only by contacts between people, and tis community is closed to interactions wit te outside world. Secondly, te maximum effect of te infected fraction on te infection rate is. Tis means tat, te infection rate cannot be iger tan te susceptible population. It is important to note tat tis claim does not imply tat te end point will be (,). Wen all or almost all people are infected, te effect of infected on infection rate may be smaller, and may even decline to zero. 4

5 3. APPLICATION OF HEARNE S METHOD TO THE ALTERNATIVE SIR MODEL In tis section, te metod of Jon Hearne will be explained, and ten te results obtained by using it on te abovementioned SIR model will be presented. 3.. Hearne s metod for function sensitivity analysis Te basis of Hearne s metod for function sensitivity analysis is te multiplication of a model function wit anoter function of specific form but variable parameterization in order to distort te model function. In is study, Hearne used triangular functions as distortion functions. Tecnically speaking, tese triangular functions ave te analytical and grapical forms sown below. For a model function r(y) defined on te interval [a, b], perturbation function (y, p, m) were m and p are parameters is given by: Here, it is important to note tat m stands for te maximum deviation from, wereas p is te point were tis deviation occurs. Since tis study does not aim at exploring te effect of different end points of model functions on te beavior, in tis paper te distortion function is formulated in suc a way tat no end point distortion occurs. Below, Figure 4 sows an example table function wic as an exponentially growing sape. If tis table function could ave an s-sape, almost linear or a less steep form, as R, R2 and R3 in Figures 6, 8 and respectively, ten suc sapes can be obtained by multiplying te original table function wit te distortion functions sown in Figures 5, 7 and 9. Wit eac possible combination of parameter values of te distortion function, a different table function sape can be obtained and a large span of possibilities can be explored.,6,4,2,8,6 r,4,2,,2,3,4,5,6,7,8,9 Figure 4: An example of a table function, r 5

6 4,5 4 3,5 3 2,5 2,5,5 Figure 5: Distortion Function wit m=,8 and p=,6 Figure 6: R, te distorted form of te table function r, distorted by 4,5 4 3,5 3 2,5 2,5,5,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9 Figure 7: Distortion Function 2 wit m= and p=,5 Figure 8: R2, te distorted form of te table function r, distorted by 2,2,8,6,4,2,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9 2 3 Figure 9: Distortion Function 3 wit m=-,64 and p=,8 Figure : R3, te distorted form of te table function r, distorted by 3,8,6,4,2,8,6,4,2,6,4,2,8,6,4,2,6,4,2,8,6,4,2,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9 R r R2 r R3 r To find te minimum distortion to te model function wic causes te undesired model beavior, Hearne formulates te problem as follows: In te solution (m *, p * ) to tis optimization problem, m * represents te magnitude of te distortion, and p * is te point at wic te function manifests most sensitivity to te canges. 6

7 Tis formulation provides te reason for terming tis approac an automated metod. Instead of making manual canges to te function, te sape of te function wic gives rise to undesired beavior can easily be found wit te aid of an optimization tool Experiments wit te original approac For te modified SIR model, tree types of model beavior are specified as te criteria by wic sensitivity will be determined. If tese beaviors are generated by minor canges in te table function, ten te model will be said to be igly sensitive to te sape of te table function. Te tree criteria and te rationale for coosing eac of tem are described subsequently. i. Maximum number of Infected people greater tan 3 in two monts In te original SIR model, te infected population reaces its maximum value of around 3 in te second week; wereas te peak value of around 2 in te alternative model wic includes te table function. Te table function sape wic will yield a similar maximum for infected people is sougt by defining te criterion as te maximum number of infected people sould exceed 3 witin two monts. ii. Maximum number of Infected people greater tan 6 in two monts Given te duration of infectivity specified in te model, recovery is quite rapid and te maximum number of people infected at any time over a two mont period is not very ig. Te ambitious criterion of exceeding 6 infected people in two monts is cosen to figure out if any cange in te table function can cause suc extreme beavior. iii. Number of susceptible people still greater tan 99 after year Tis model is intended to model te dynamics of an epidemic. Terefore, if tere is no infection occurring, te model cannot be considered to acieve its purpose. Tis criterion is cosen in an attempt to understand if te sape of te table function could cause te model to be invalid. Tree optimization problems, one for eac of tese criteria, are formulated as proposed by Hearne. An additional constraint is added to ensure tat te values of te table function lie between and, preventing infection rates tat are negative or iger tan te susceptible population. Teoretically, m can take any value; owever, since te table function values cannot exceed, or go below, it is unnecessary to searc for m in a large interval. Terefore, m values are assumed to be between -5 and 5. Te matematical models specific to eac optimization problem are included in Appendix I Results To find te function sapes wic create undesired model beaviors, te optimization problems described above ave been solved using te optimization tool of Powersim Studio 7. Optimal solutions are presented in Table. 7

8 Table : Results of applying Hearne's metod to te alternative SIR model CRITERION CRITERION 2 CRITERION 3 m,69 4,695 -,47 p,,238,,2,8,6,4,2,,2,3,4,5,6,7,8,9 2,5 2,5,5,,2,3,4,5,6,7,8,9 Figure 2: Table function distorted to meet criterion 2 Figure : Table function distorted to meet criterion,2,8,6,4,2,,2,3,4,5,6,7,8,9 Figure 3: Table function distorted to meet criterion 3 Figures, 2 and 3 sow ow te original table function is distorted to meet eac criterion wit te function defined by te respective parameters in Table. In Figure, it can be seen tat if te table function is canged to ave a convex sape rater tan a linearly increasing one, it ten causes te maximum infected population to be greater tan 3. Te relationsip depicted in Figure 2 means tat to ave maximum infected population greater tan 6, infection rate sould be more tan, even twice as muc as, te susceptible population wen te infected people are between 5-% of te total population. In Figure 3, we see tat infection does not occur and te susceptible population remains ig if te table function values are sligtly decreased. Te results generally indicate tat te model beavior is more sensitive to te table function in terms of te tird criterion, because te absolute maximum deviation from unity (m) is smaller in tat case. A feasible solution for te second optimization problem couldn t be found; instead, te result obtained after removing te constraint on te table function is given. Since suc a distortion does not ave a real life meaning, it can be said tat te beavior is insensitive to te table function in terms of criterion Discussion on te observations and identifying potential improvements Te undesired beavior in te first and tird cases was obtained by a distortion of te table function made using a triangular function. However, tis does not mean tat tese are te only function sape canges wic can generate te undesired beavior. For example, an s-saped pattern wic requires te original function to increase initially ten to decrease, or vice versa, could potentially ave created te same undesired beavior. As Hearne pointed out, triangular functions are not capable of making different canges on te model functions in two different intervals. Wit teir end 8

9 points at, tey cannot decrease te function on one side, wile increasing it on te oter. However, in order to treat uncertainty in functional form torougly, suc variations in perturbations are necessary. Terefore, an improvement to Hearne s metod lies in determining alternative forms of distortion functions tat are able to create suc two-sided distortions. Furtermore, it is observed tat tis metod unnecessarily makes distortions on te regions of table function wic are not used during te simulation. For te first criterion, te segment after.3 is not used, because te igest infected population is 3. Similarly, only te part before. is used for te tird criterion. Tis indicates tat te distortions after tose points proved unnecessary, but ad to be made owing to te sape of te distortion function. A furter improvement lies in avoiding tese unnecessary distortions. Tere is anoter reason for using suc functions for te distortion of table functions: Tis table function ends at (, ) and single-extreme triangular functions wit stable end points at can be used to perturb tem. On te oter and, for table functions passing troug (, ) and going furter, distorted functions sould also pass troug (, ) and to ensure tis, distortion function as to pass troug (, ) as well. It is not possible to acieve tis by a partial function wit two segments wose end points are. Terefore, it is inevitable to use functions wic can provide different distortions in different intervals. As it was said before, tis table function assumes tat te effect of infected population ratio on infection rate increases till. However, it may not be logical to expect all te susceptible people to become infected even if infected ratio is very ig. Terefore, tis table function may saturate below. Tis opinion in fact points out te general fact tat end points of table functions may also be subject to uncertainty. Also, variable end points increase te variety of function sapes tat can be obtained after perturbation. Hence, tis function sensitivity analysis metod sould include te possible variations at te end points as well. 4. EXTENSIONS TO HEARNE S METHOD In tis section, te formulation and implementation of te proposed extensions on te model will be explained. Ten, te results of te implementation will be described. 4.. Extensions 4... Double-extreme triangular functions In selecting distortion functions tere is a trade-off between teir ability to provide te desired variety of perturbations and te number of parameters tey ave. Te variety of perturbations is important for te reasons previously described. As for te number of parameters, te iger te number, te longer it takes to solve te optimization problem and te more difficult it is to interpret te impact of te parameters on te sape of te function. Considering tis trade-off, te simplest extension tat can be made to triangular functions in order to obtain two-sided distortions is to use double-extreme triangular functions, tat is piece-wise functions wit two extreme points instead of one. Suc functions can be defined using four parameters wit meanings similar to te parameters 9

10 of triangular functions, and tey yield two-sided distortions wen m and m 2 ave opposite signs. If tey ave te same sign, ten tey just increase te variety. Te definition and an example of tese functions, for a model function r(x) defined in interval [, ] are given below (see Figures 4 and 5): 3 2,5 2,5,5 Double Single,,2,3,4,5,6,7,8,9 Figure 4: An example of a double-extreme triangular distortion function wit m=.2, p=.2, m2=-.3 and p2=.7 ( Double is in Figure 5),8,6,4,2,8,6,4,2 R Double R Single,,2,3,4,5,6,7,8,9 Figure 5: Te function in Figure 4 distorted by te single and double-extreme functions in Figure 4 r Variable end points Wen te end points are allowed to vary, te distortion functions defined above will ave to be modified as follows, were l and u are te values of te table function at and, respectively: For single extreme triangular functions: For double extreme triangular functions:

11 It is wortwile to mention tat in case of variable end points, parameter m represents te maximum deviation from te lower or te closer end point, not te deviation from unity anymore Experiments wit te extended metod Te experimental setup for implementing te extensions is te same as explained in section 3.2. An optimization problem for eac criterion and an extension is defined. For double-extreme distortion functions, bot te cases in wic te end points are stable and in wic tey are variable are considered. Because te population is assumed isolated from external infection sources (see section 2), even te distorted table function is assumed to start at (,) and no cange is made to tis point. Yet, te lower end point of te distortion function is still assumed to be a variable and take values different from, because te beginning point also affects te sape of te function. Te matematical models specific to eac optimization problem are detailed in Appendix I. To ensure two-sided distortion using double triangular functions, wic was te reason for introducing tem in te first place, m and m 2 can be forced to ave opposite signs Results Te optimization problems are solved as before. Te optimal solution of eac problem can be seen in Table 2 below: Table 2: Results of applying Hearne's metod wit extensions to te alternative SIR model CRITERION CRITERION 2 CRITERION 3 SINGLE- EXTREME DISTORTION FUNCTION DOUBLE- EXTREME DISTORTION FUNCTION END POINTS STABLE AT (Original metod) VARIABLE END POINTS END POINTS STABLE AT VARIABLE END POINTS m,69 4,695 -,47 p,,238, m,725 4,8, p,,364,55 l,,,295 u,874,69,428 m,74 4,745 -,446 p,,42, m2 -,5,,2 p2,,,366 m 4,47 2,383, p,484,,23 m2 2,257 2,782, p2,957,729,556 l,736,,3 u,3,,869

12 Tere is no feasible solution wic makes te maximum infected population greater tan 6 in any of te cases. Te solutions above (in bold typeface) belong to te modified optimization problem witout te constraint on te values of te distorted table function. Te graps of te distortion functions wit tese parameters and ow tey perturb te table function can be seen in Appendix II. In te context of te modified SIR model, te results above, supported by te visualizations in Appendix II, are very similar to te results of te original metod. Tey indicate tat te model is moderately sensitive to canges in te sape of table function f wen te sensitivity criterion is tat te maximum value of infected population exceeds 3% of te total population. If tis tresold is increased, tat is, if te model is expected to demonstrate even less desirable beavior, ten it can be said to be totally insensitive, because te distortions required to produce suc beavior no longer ave real life interpretations. However, if te undesired beavior is tat te final susceptible population is greater tan 99% of te total population, ten te beavior can be said to be igly sensitive to te canges in te sape of f, because sligt decreases at te initial points of te table function are sufficient to keep te infection rate very low. In te context of metod extension, it is seen tat on tis model and for tese criteria, te single-extreme triangular function wit stable end points was sufficient to carry out te function sensitivity analyses. Te double-extreme functions cosen by te optimization algoritm act very similarly to te single-extreme ones, because two of teir four intervals are negligibly narrow. Wen te end points are allowed to vary, tey are cosen different from, but tese canges are not really effective, because tese parts of te table function are not used. Still, te extensions proposed ere are teoretically necessary because tey enlarge te solution space and so facilitate a fuller consideration of uncertainty. Furtermore, tese experiments are useful in determining te utility of te metod, delineating necessary improvements and demonstrating teir effects Discussion on te observations For some of te optimal solutions, te end point deviates from unity. However, tis final segment of te table function is never even used. A u value lower tan one is cosen by te optimization algoritm eiter arbitrarily since it as no effect, or due to te ease of creating lower values for oter parts of te function wit a small end point. Terefore, before deriving te conclusion tat te beavior is sensitive to te canges in te end point of te table function if it falls below u, one needs to verify weter tis part of te function is used or weter tere is a significant effect on te part actually used. Furter, te variable end points add anoter level of complication to te problem. It is indeed straigtforward to include tem in te problem and tey are expected to stay at if variations don t cause a significant cange in te model beavior. However, as in te case of te double-extreme distortion function and te first criterion, altoug it is known tat tere is a feasible solution wen te end points are, te searc algoritm does not find tis. To overcome tis difficulty, eiter te searc process can be extended, wic brings te dilemma of quality versus time and effort; or anoter optimization tool can be used, wic decreases te efficiency of te analysis procedure by separating te environments in wic te optimization and simulation are executed. 2

13 Moreover, altoug double-extreme triangular functions are able to generate two-sided distortions, teir perturbation variety is still limited. To increase tis variety, furter functional forms, suc as cubic polynomials or sinusoidal functions, can be tried out; and te added value of aving tis variety against te complications and difficulty of interpretation caused by te iger number of parameters, can be assessed. 5. DISCUSSION AND CONCLUSION In tis study, te applicability of Hearne s automated sensitivity analysis metod for system dynamics functions as been tested by applying te metod to a simple model containing a table function. Extensions to te metod ave been developed and implemented. Specific sortcomings of bot te original metod and te extended version are delineated in sections 3.4 and 4.4. In addition to tese, tere are some general issues tat need to be addressed. First, te claims made about te table function s range in use sould be generalized, because it is not unusual for models to use only a limited part of te table functions under given initial conditions and parameter values. However, it is not possible to capture te exact end points of te domain interval used in runtime, and to make distortions only in tis interval. Terefore, it is necessary to ceck te active domain of te table functions, and ow muc of te distortions are really used in te simulation before deriving conclusions from te results obtained wit tis metod. Second, in tis study, undesired beaviors were generated by single-extreme functions in only one direction. To fully understand te added value of double-extreme functions, different undesired beavior types sould be determined and tested. Also, one may argue tat sensitivity in tis study is defined more in terms of numerical canges ere, rater tan in terms of significant beavioral canges. Tis claim is undeniably true; but wit an optimization tool in wic te constraints can be applied only once during te simulation, not continuously, it is not straigtforward to capture beavioral canges. Indeed, automated beavioral cange analysis metods suc as pattern recognition would need to be incorporated in te extended Hearne metod if te goal of creating an automated sensitivity analysis metod for model functions is to be attained. Tird, in tis paper te metod is applied on a simple model wit a single table function, and undesired beaviors are selected on te basis of te beavior of a single stock variable. If te metod is used on a larger model to create simultaneous distortions on multiple table functions and criteria dependent on combinations of several stock variables are defined, ten te value of tis automated metod can be establised. Bot te original and te extended version of Hearne s function sensitivity analysis metod were applied using te optimization tools available in system dynamics software, demonstrating te promise of te metod to deliver automated sensitivity analysis of functions in system dynamics models. A number of points wic need attention if tis metod is to be used routinely for table functions sensitivity analysis provide te points of departure in future work on tis issue. 3

14 REFERENCES Clemson, Barry, Yongming Tang, James Pyne, and Resit Unal. "Efficient metods for sensitivity analysis." System Dynamics Review, no. (995). Ford, A, J.S. Amlin, and G. Backus. "A Practical Approac to Sensitivity Testing of System Dynamics Models." Proceedings of te 983 International System Dynamics Conference. Cestnut Hill, MA, 983: Ford, A., and H Flynn. "Statistical Screening of System Dynamics Models." System Dynamics Review, 25. Hearne, J, W. "An Automated Metod for Extending Sensitivity Analysis to Model Functions." Natural Resource Modleing, 2. Hearne, J.W. "An approac to resolving parameter sensitivity problem is system dynamics metodology." Applied Matematical Modelling, 987. Hekimoglu, Mustafa, and Yaman Barlas. "Sensitivity Analysis of System Dynamics Models by Beavior Pattern Measures." Proceedings of te 28t International Conference of te System Dynamics Society. 2. Kleijnen, J.P.C. "Sensitivity analysis and optimization of system dynamics models: regression analysis and statistical design of experiments." System Dynamics Review, 995. Meadows, D. "Te Unavoidable a Priori." In Elements of te System Dynamics Metod, by Jørgen Randers. Cambridge, Massacusetts, USA: Productivity Press, 98. Miller, J.H. "Active Nonlinear Tests (ANTs) of Complex Simulation Models." Management Science, 998. Sterman, J. Business Dynamics: Systems Tinking and Modeling for a Complex World. Irwin/McGraw-Hill, 2. Tank-Nielsen, T. "Sensitivity Analysis in System Dynamics." In Elements of te System Dynamics Metod, by Jørgen Randers. Cambridge, Massacusetts, USA: Productivity Press, 98. Yucel, Gonenc, and Yaman Barlas. "Automated Parameter Specification in Dynamic Feedback Models Based on Beavior Pattern Features." System Dynamics Review 27, no. 2 (2):

15 APPENDIX I: OPTIMIZATION PROBLEMS. Single extreme, stable end points, model beavior () i s.t. (2) x (3) (4) (5) were InfectionFunction[i] = InfectionFunction[i] * [i] InfectionFunction[] = {,.9,.2,.29,.46,.5,.66,.69,.794,.96, } were () Objective function: Te optimization problem for sensitivity analysis is stated s fi di g te i i u distortio wic cre tes te u desired odel be vior. Terefore, te distortio is e sured by te squ re of xi u deviation of te distortion function from unity, to ensure tat absolute distortion is taken into account. (2) Undesired model beavior constraint: Te first undesired model beavior criterion is formulated as a constraint wic restricts te maximum value of te Infected population over te simulation time to be greater tan 3. In Powersim, tis constraint was defined by using RUNMAX() function. (3) Feasibility constraint: It is known tat te values below and above ave no real life meaning for te table function used in tis study. However, it is not possible to set a constraint to ceck te distorted values at eac time point. Terefore, an array of discrete values of te distorted table function, namely InfectionFunction[], is created and restricted to te interval of [, ], altoug it is known tat te distortions are continuous. Eac element of InfectionFunction[] is formed by multiplying eac element of te original distortion function array (InfectionFunction[]) wit te corresponding value of te distortion function([i]). (4) Searc range of p: As it can be recalled, parameter p represents te point were te maximum distortion from unity occurs. Terefore, it as to witin te domain of table function, wic is [, ] in tis case. (5) Searc range of m: Tere is no limit on te values of m, but it is anticipated tat ig values cause infeasible distortions. Terefore, te searc range is specified as [-5, 5] to include downward distortions as well. 2. Single extreme, stable end points, model beavior 2 Problem, except te first constraint wic is replaced by 3. Single extreme, stable end points, model beavior 3 Problem, except te first constraint wic is replaced by 5

16 4. Single extreme, variable end points, model beavior () s.t. (2) (3) (4) (5) (6) (7) were InfectionFunction[i] = InfectionFunction[i] * [i] InfectionFunction[] = {,.9,.2,.29,.46,.5,.66,.69,.794,.96, } (6) Searc range of l: In addition to te problem, searc ranges for te end point parameters are specified in tis case. For l, te end point at, tis range is [, 2] because it is tougt tat te values iger tan 2 would easily cause infeasibility. (7) Searc range of u: Te end point of te table function at is already, and since multiplying it wit values iger tan would cause infeasible values, te searc range of u is kept between and. 5. Single extreme, variable end points, model beavior 2 Problem 4, except te first constraint wic is replaced by 6. Single extreme, variable end points, model beavior 3 Problem 4, except te first constraint wic is replaced by 7. Double extreme, stable end points, model beavior () (2) (3) (4) (5) (6) were s.t. InfectionFunction[i] = InfectionFunction[i] * [i] 6

17 InfectionFunction[] = {,.9,.2,.29,.46,.5,.66,.69,.794,.96, } () Objective function: Tis time, two parameters, bot m and m 2 contribute to te maximum distortion, ence tey are bot included in te objective function. (4) Searc range of p and p 2: Te searc range [,] applies to bot p parameters. (5) Searc range of m and m 2: As specified in te first problem, te searc range [- 5, 5] applies to bot m parameters. (6) Difference of p s: Te distortion function is defined based on te assumption tat p precedes p 2. Wit tis constraint, it is guaranteed tat p is smaller tan p Double extreme, stable end points, model beavior 2 Problem 7, except te first constraint wic is replaced by 9. Double extreme, stable end points, model beavior 3 Problem 5, except te first constraint wic is replaced by. Double extreme, variable end points, model beavior () i s.t. (2) x (3) (4) (5) (6) (7) (8) were InfectionFunction[i] = InfectionFunction[i] * [i] InfectionFunction[] = {,.9,.2,.29,.46,.5,.66,.69,.794,.96, } 7

18 (7) and (8) Searc ranges of end points: Te searc range of bot of te end points is specified as te interval [,] because te values out of tis interval are infeasible.. Double extreme, variable end points, model beavior 2 Problem, except te first constraint wic is replaced by 2. Double extreme, variable end points, model beavior 3 Problem, except te first constraint wic is replaced by 8

19 ,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9 APPENDIX II: EXPERIMENT RESULTS. Single-Extreme Triangular Distortion Function a. Stable End Points i. Criterion (Model beavior: Max(I)>3) m,69 p,,8,6,4,2,,8,6,4,2,,2,8,6,4,2,,2,3,4,5,6,7,8,9 6, ii. Criterion 2 (Model beavior: Max(I)>6 ) m 4,695 p,238 2,5 5, 2 4, 3, 2,,,,5,5,,2,3,4,5,6,7,8,9 iii. Criterion 3 (Model beavior: Final S>99) m -,47 p,,2,2,,8,6,4,2,,8,6,4,2,,2,3,4,5,6,7,8,9 9

20 ,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9 b. Variable End Points i. Criterion (Model beavior: Max(I)>3) m,725 p, l, u,874,8,6,4,2,,8,6,4,2,,2,8,6,4,2,,2,3,4,5,6,7,8,9 ii. Criterion 2 (Model beavior: Max(I)>6 ) m 4,8 p,364 l, u,69 6, 3 5, 2,5 4, 3, 2,,, 2,5,5,,2,3,4,5,6,7,8,9 iii. Criterion 3 (Model beavior: Final S>99) m, p,55 l,295 u,428,45,4,35,3,25,2,5,,5,,2,8,6,4,2,,2,3,4,5,6,7,8,9 2

21 ,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9 2. Double-Extreme Triangular Distortion Function a. Stable End Points i. Criterion (Model beavior: Max(I)>3) m,74 p, m2 -,5 p2,,8,6,4,2,,8,6,4,2,,2,8,6,4,2,,2,3,4,5,6,7,8,9 ii. Criterion 2 (Model beavior: Max(I)>6 ) m 4,745 p,42 m2, p2, 7, 3 6, 5, 4, 3, 2,, 2,5 2,5,5,,,2,3,4,5,6,7,8,9 iii. Criterion 3 (Model beavior: Final S>99) m -,446 p, m2,2 p2,366,2,2,,8,8,6,4,2,6,4,2,,,2,3,4,5,6,7,8,9 2

22 ,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9 b. Variable End Points i. Criterion (Model beavior: Max(I)>3) m 4,47 p,484 m2 2,257 p2,957 l,736 u,3 2,,2,5,,5,8,6,4,2 Origin al Distor ted,,,2,3,4,5,6,7,8,9,,2,3,4,5,6,7,8,9 ii. Criterion 2 (Model beavior: Max(I)>6 ) m 2,383 p, m2 2,782 p2,729 l, u, 4, 3,5 3, 2,5 2,,5,,5, 3 2,5 2,5,5,,2,3,4,5,6,7,8,9 iii. Criterion 3 (Model beavior: Final S>99) m, p,23 m2, p2,556 l,3 u,869,,2,8,6,4,2,8,6,4,2,,,2,3,4,5,6,7,8,9 22

4.1 Tangent Lines. y 2 y 1 = y 2 y 1

4.1 Tangent Lines. y 2 y 1 = y 2 y 1 41 Tangent Lines Introduction Recall tat te slope of a line tells us ow fast te line rises or falls Given distinct points (x 1, y 1 ) and (x 2, y 2 ), te slope of te line troug tese two points is cange

More information

Bounding Tree Cover Number and Positive Semidefinite Zero Forcing Number

Bounding Tree Cover Number and Positive Semidefinite Zero Forcing Number Bounding Tree Cover Number and Positive Semidefinite Zero Forcing Number Sofia Burille Mentor: Micael Natanson September 15, 2014 Abstract Given a grap, G, wit a set of vertices, v, and edges, various

More information

Fast Calculation of Thermodynamic Properties of Water and Steam in Process Modelling using Spline Interpolation

Fast Calculation of Thermodynamic Properties of Water and Steam in Process Modelling using Spline Interpolation P R E P R N T CPWS XV Berlin, September 8, 008 Fast Calculation of Termodynamic Properties of Water and Steam in Process Modelling using Spline nterpolation Mattias Kunick a, Hans-Joacim Kretzscmar a,

More information

MATH 5a Spring 2018 READING ASSIGNMENTS FOR CHAPTER 2

MATH 5a Spring 2018 READING ASSIGNMENTS FOR CHAPTER 2 MATH 5a Spring 2018 READING ASSIGNMENTS FOR CHAPTER 2 Note: Tere will be a very sort online reading quiz (WebWork) on eac reading assignment due one our before class on its due date. Due dates can be found

More information

Linear Interpolating Splines

Linear Interpolating Splines Jim Lambers MAT 772 Fall Semester 2010-11 Lecture 17 Notes Tese notes correspond to Sections 112, 11, and 114 in te text Linear Interpolating Splines We ave seen tat ig-degree polynomial interpolation

More information

More on Functions and Their Graphs

More on Functions and Their Graphs More on Functions and Teir Graps Difference Quotient ( + ) ( ) f a f a is known as te difference quotient and is used exclusively wit functions. Te objective to keep in mind is to factor te appearing in

More information

Piecewise Polynomial Interpolation, cont d

Piecewise Polynomial Interpolation, cont d Jim Lambers MAT 460/560 Fall Semester 2009-0 Lecture 2 Notes Tese notes correspond to Section 4 in te text Piecewise Polynomial Interpolation, cont d Constructing Cubic Splines, cont d Having determined

More information

Our Calibrated Model has No Predictive Value: An Example from the Petroleum Industry

Our Calibrated Model has No Predictive Value: An Example from the Petroleum Industry Our Calibrated Model as No Predictive Value: An Example from te Petroleum Industry J.N. Carter a, P.J. Ballester a, Z. Tavassoli a and P.R. King a a Department of Eart Sciences and Engineering, Imperial

More information

2.8 The derivative as a function

2.8 The derivative as a function CHAPTER 2. LIMITS 56 2.8 Te derivative as a function Definition. Te derivative of f(x) istefunction f (x) defined as follows f f(x + ) f(x) (x). 0 Note: tis differs from te definition in section 2.7 in

More information

4.2 The Derivative. f(x + h) f(x) lim

4.2 The Derivative. f(x + h) f(x) lim 4.2 Te Derivative Introduction In te previous section, it was sown tat if a function f as a nonvertical tangent line at a point (x, f(x)), ten its slope is given by te it f(x + ) f(x). (*) Tis is potentially

More information

Comparison of the Efficiency of the Various Algorithms in Stratified Sampling when the Initial Solutions are Determined with Geometric Method

Comparison of the Efficiency of the Various Algorithms in Stratified Sampling when the Initial Solutions are Determined with Geometric Method International Journal of Statistics and Applications 0, (): -0 DOI: 0.9/j.statistics.000.0 Comparison of te Efficiency of te Various Algoritms in Stratified Sampling wen te Initial Solutions are Determined

More information

An Algorithm for Loopless Deflection in Photonic Packet-Switched Networks

An Algorithm for Loopless Deflection in Photonic Packet-Switched Networks An Algoritm for Loopless Deflection in Potonic Packet-Switced Networks Jason P. Jue Center for Advanced Telecommunications Systems and Services Te University of Texas at Dallas Ricardson, TX 75083-0688

More information

2 The Derivative. 2.0 Introduction to Derivatives. Slopes of Tangent Lines: Graphically

2 The Derivative. 2.0 Introduction to Derivatives. Slopes of Tangent Lines: Graphically 2 Te Derivative Te two previous capters ave laid te foundation for te study of calculus. Tey provided a review of some material you will need and started to empasize te various ways we will view and use

More information

Multi-Stack Boundary Labeling Problems

Multi-Stack Boundary Labeling Problems Multi-Stack Boundary Labeling Problems Micael A. Bekos 1, Micael Kaufmann 2, Katerina Potika 1 Antonios Symvonis 1 1 National Tecnical University of Atens, Scool of Applied Matematical & Pysical Sciences,

More information

Vector Processing Contours

Vector Processing Contours Vector Processing Contours Andrey Kirsanov Department of Automation and Control Processes MAMI Moscow State Tecnical University Moscow, Russia AndKirsanov@yandex.ru A.Vavilin and K-H. Jo Department of

More information

The Euler and trapezoidal stencils to solve d d x y x = f x, y x

The Euler and trapezoidal stencils to solve d d x y x = f x, y x restart; Te Euler and trapezoidal stencils to solve d d x y x = y x Te purpose of tis workseet is to derive te tree simplest numerical stencils to solve te first order d equation y x d x = y x, and study

More information

Two Modifications of Weight Calculation of the Non-Local Means Denoising Method

Two Modifications of Weight Calculation of the Non-Local Means Denoising Method Engineering, 2013, 5, 522-526 ttp://dx.doi.org/10.4236/eng.2013.510b107 Publised Online October 2013 (ttp://www.scirp.org/journal/eng) Two Modifications of Weigt Calculation of te Non-Local Means Denoising

More information

Cubic smoothing spline

Cubic smoothing spline Cubic smooting spline Menu: QCExpert Regression Cubic spline e module Cubic Spline is used to fit any functional regression curve troug data wit one independent variable x and one dependent random variable

More information

Haar Transform CS 430 Denbigh Starkey

Haar Transform CS 430 Denbigh Starkey Haar Transform CS Denbig Starkey. Background. Computing te transform. Restoring te original image from te transform 7. Producing te transform matrix 8 5. Using Haar for lossless compression 6. Using Haar

More information

Density Estimation Over Data Stream

Density Estimation Over Data Stream Density Estimation Over Data Stream Aoying Zou Dept. of Computer Science, Fudan University 22 Handan Rd. Sangai, 2433, P.R. Cina ayzou@fudan.edu.cn Ziyuan Cai Dept. of Computer Science, Fudan University

More information

The (, D) and (, N) problems in double-step digraphs with unilateral distance

The (, D) and (, N) problems in double-step digraphs with unilateral distance Electronic Journal of Grap Teory and Applications () (), Te (, D) and (, N) problems in double-step digraps wit unilateral distance C Dalfó, MA Fiol Departament de Matemàtica Aplicada IV Universitat Politècnica

More information

The impact of simplified UNBab mapping function on GPS tropospheric delay

The impact of simplified UNBab mapping function on GPS tropospheric delay Te impact of simplified UNBab mapping function on GPS troposperic delay Hamza Sakidin, Tay Coo Cuan, and Asmala Amad Citation: AIP Conference Proceedings 1621, 363 (2014); doi: 10.1063/1.4898493 View online:

More information

3.6 Directional Derivatives and the Gradient Vector

3.6 Directional Derivatives and the Gradient Vector 288 CHAPTER 3. FUNCTIONS OF SEVERAL VARIABLES 3.6 Directional Derivatives and te Gradient Vector 3.6.1 Functions of two Variables Directional Derivatives Let us first quickly review, one more time, te

More information

Numerical Derivatives

Numerical Derivatives Lab 15 Numerical Derivatives Lab Objective: Understand and implement finite difference approximations of te derivative in single and multiple dimensions. Evaluate te accuracy of tese approximations. Ten

More information

Unsupervised Learning for Hierarchical Clustering Using Statistical Information

Unsupervised Learning for Hierarchical Clustering Using Statistical Information Unsupervised Learning for Hierarcical Clustering Using Statistical Information Masaru Okamoto, Nan Bu, and Tosio Tsuji Department of Artificial Complex System Engineering Hirosima University Kagamiyama

More information

12.2 TECHNIQUES FOR EVALUATING LIMITS

12.2 TECHNIQUES FOR EVALUATING LIMITS Section Tecniques for Evaluating Limits 86 TECHNIQUES FOR EVALUATING LIMITS Wat ou sould learn Use te dividing out tecnique to evaluate its of functions Use te rationalizing tecnique to evaluate its of

More information

Section 1.2 The Slope of a Tangent

Section 1.2 The Slope of a Tangent Section 1.2 Te Slope of a Tangent You are familiar wit te concept of a tangent to a curve. Wat geometric interpretation can be given to a tangent to te grap of a function at a point? A tangent is te straigt

More information

, 1 1, A complex fraction is a quotient of rational expressions (including their sums) that result

, 1 1, A complex fraction is a quotient of rational expressions (including their sums) that result RT. Complex Fractions Wen working wit algebraic expressions, sometimes we come across needing to simplify expressions like tese: xx 9 xx +, xx + xx + xx, yy xx + xx + +, aa Simplifying Complex Fractions

More information

Symmetric Tree Replication Protocol for Efficient Distributed Storage System*

Symmetric Tree Replication Protocol for Efficient Distributed Storage System* ymmetric Tree Replication Protocol for Efficient Distributed torage ystem* ung Cune Coi 1, Hee Yong Youn 1, and Joong up Coi 2 1 cool of Information and Communications Engineering ungkyunkwan University

More information

Multi-Objective Particle Swarm Optimizers: A Survey of the State-of-the-Art

Multi-Objective Particle Swarm Optimizers: A Survey of the State-of-the-Art Multi-Objective Particle Swarm Optimizers: A Survey of te State-of-te-Art Margarita Reyes-Sierra and Carlos A. Coello Coello CINVESTAV-IPN (Evolutionary Computation Group) Electrical Engineering Department,

More information

1.4 RATIONAL EXPRESSIONS

1.4 RATIONAL EXPRESSIONS 6 CHAPTER Fundamentals.4 RATIONAL EXPRESSIONS Te Domain of an Algebraic Epression Simplifying Rational Epressions Multiplying and Dividing Rational Epressions Adding and Subtracting Rational Epressions

More information

ANTENNA SPHERICAL COORDINATE SYSTEMS AND THEIR APPLICATION IN COMBINING RESULTS FROM DIFFERENT ANTENNA ORIENTATIONS

ANTENNA SPHERICAL COORDINATE SYSTEMS AND THEIR APPLICATION IN COMBINING RESULTS FROM DIFFERENT ANTENNA ORIENTATIONS NTNN SPHRICL COORDINT SSTMS ND THIR PPLICTION IN COMBINING RSULTS FROM DIFFRNT NTNN ORINTTIONS llen C. Newell, Greg Hindman Nearfield Systems Incorporated 133. 223 rd St. Bldg. 524 Carson, C 9745 US BSTRCT

More information

A Cost Model for Distributed Shared Memory. Using Competitive Update. Jai-Hoon Kim Nitin H. Vaidya. Department of Computer Science

A Cost Model for Distributed Shared Memory. Using Competitive Update. Jai-Hoon Kim Nitin H. Vaidya. Department of Computer Science A Cost Model for Distributed Sared Memory Using Competitive Update Jai-Hoon Kim Nitin H. Vaidya Department of Computer Science Texas A&M University College Station, Texas, 77843-3112, USA E-mail: fjkim,vaidyag@cs.tamu.edu

More information

CESILA: Communication Circle External Square Intersection-Based WSN Localization Algorithm

CESILA: Communication Circle External Square Intersection-Based WSN Localization Algorithm Sensors & Transducers 2013 by IFSA ttp://www.sensorsportal.com CESILA: Communication Circle External Square Intersection-Based WSN Localization Algoritm Sun Hongyu, Fang Ziyi, Qu Guannan College of Computer

More information

Effect of GPS Tropospheric Delay Neill Mapping Function Simplification

Effect of GPS Tropospheric Delay Neill Mapping Function Simplification Malaysian Journal of Matematical Sciences 3(1): 95-107 (2009) Effect of GPS Troposperic Delay Neill Mapping Function Simplification 1 Hamza Sakidin, 1 Mod Rizam Abu Bakar, 2 Abdul Rasid Moamed Sariff,

More information

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Spring

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Spring Has-Based Indexes Capter 11 Comp 521 Files and Databases Spring 2010 1 Introduction As for any index, 3 alternatives for data entries k*: Data record wit key value k

More information

13.5 DIRECTIONAL DERIVATIVES and the GRADIENT VECTOR

13.5 DIRECTIONAL DERIVATIVES and the GRADIENT VECTOR 13.5 Directional Derivatives and te Gradient Vector Contemporary Calculus 1 13.5 DIRECTIONAL DERIVATIVES and te GRADIENT VECTOR Directional Derivatives In Section 13.3 te partial derivatives f x and f

More information

Utilizing Call Admission Control to Derive Optimal Pricing of Multiple Service Classes in Wireless Cellular Networks

Utilizing Call Admission Control to Derive Optimal Pricing of Multiple Service Classes in Wireless Cellular Networks Utilizing Call Admission Control to Derive Optimal Pricing of Multiple Service Classes in Wireless Cellular Networks Okan Yilmaz and Ing-Ray Cen Computer Science Department Virginia Tec {oyilmaz, ircen}@vt.edu

More information

An Analytical Approach to Real-Time Misbehavior Detection in IEEE Based Wireless Networks

An Analytical Approach to Real-Time Misbehavior Detection in IEEE Based Wireless Networks Tis paper was presented as part of te main tecnical program at IEEE INFOCOM 20 An Analytical Approac to Real-Time Misbeavior Detection in IEEE 802. Based Wireless Networks Jin Tang, Yu Ceng Electrical

More information

Redundancy Awareness in SQL Queries

Redundancy Awareness in SQL Queries Redundancy Awareness in QL Queries Bin ao and Antonio Badia omputer Engineering and omputer cience Department University of Louisville bin.cao,abadia @louisville.edu Abstract In tis paper, we study QL

More information

On the Use of Radio Resource Tests in Wireless ad hoc Networks

On the Use of Radio Resource Tests in Wireless ad hoc Networks Tecnical Report RT/29/2009 On te Use of Radio Resource Tests in Wireless ad oc Networks Diogo Mónica diogo.monica@gsd.inesc-id.pt João Leitão jleitao@gsd.inesc-id.pt Luis Rodrigues ler@ist.utl.pt Carlos

More information

Tuning MAX MIN Ant System with off-line and on-line methods

Tuning MAX MIN Ant System with off-line and on-line methods Université Libre de Bruxelles Institut de Recerces Interdisciplinaires et de Développements en Intelligence Artificielle Tuning MAX MIN Ant System wit off-line and on-line metods Paola Pellegrini, Tomas

More information

Intra- and Inter-Session Network Coding in Wireless Networks

Intra- and Inter-Session Network Coding in Wireless Networks Intra- and Inter-Session Network Coding in Wireless Networks Hulya Seferoglu, Member, IEEE, Atina Markopoulou, Member, IEEE, K K Ramakrisnan, Fellow, IEEE arxiv:857v [csni] 3 Feb Abstract In tis paper,

More information

Coarticulation: An Approach for Generating Concurrent Plans in Markov Decision Processes

Coarticulation: An Approach for Generating Concurrent Plans in Markov Decision Processes Coarticulation: An Approac for Generating Concurrent Plans in Markov Decision Processes Kasayar Roanimanes kas@cs.umass.edu Sridar Maadevan maadeva@cs.umass.edu Department of Computer Science, University

More information

Distributed and Optimal Rate Allocation in Application-Layer Multicast

Distributed and Optimal Rate Allocation in Application-Layer Multicast Distributed and Optimal Rate Allocation in Application-Layer Multicast Jinyao Yan, Martin May, Bernard Plattner, Wolfgang Mülbauer Computer Engineering and Networks Laboratory, ETH Zuric, CH-8092, Switzerland

More information

12.2 Techniques for Evaluating Limits

12.2 Techniques for Evaluating Limits 335_qd /4/5 :5 PM Page 863 Section Tecniques for Evaluating Limits 863 Tecniques for Evaluating Limits Wat ou sould learn Use te dividing out tecnique to evaluate its of functions Use te rationalizing

More information

Truncated Newton-based multigrid algorithm for centroidal Voronoi diagram calculation

Truncated Newton-based multigrid algorithm for centroidal Voronoi diagram calculation NUMERICAL MATHEMATICS: Teory, Metods and Applications Numer. Mat. Teor. Met. Appl., Vol. xx, No. x, pp. 1-18 (200x) Truncated Newton-based multigrid algoritm for centroidal Voronoi diagram calculation

More information

MAC-CPTM Situations Project

MAC-CPTM Situations Project raft o not use witout permission -P ituations Project ituation 20: rea of Plane Figures Prompt teacer in a geometry class introduces formulas for te areas of parallelograms, trapezoids, and romi. e removes

More information

Chapter K. Geometric Optics. Blinn College - Physics Terry Honan

Chapter K. Geometric Optics. Blinn College - Physics Terry Honan Capter K Geometric Optics Blinn College - Pysics 2426 - Terry Honan K. - Properties of Ligt Te Speed of Ligt Te speed of ligt in a vacuum is approximately c > 3.0µ0 8 mês. Because of its most fundamental

More information

MAP MOSAICKING WITH DISSIMILAR PROJECTIONS, SPATIAL RESOLUTIONS, DATA TYPES AND NUMBER OF BANDS 1. INTRODUCTION

MAP MOSAICKING WITH DISSIMILAR PROJECTIONS, SPATIAL RESOLUTIONS, DATA TYPES AND NUMBER OF BANDS 1. INTRODUCTION MP MOSICKING WITH DISSIMILR PROJECTIONS, SPTIL RESOLUTIONS, DT TYPES ND NUMBER OF BNDS Tyler J. lumbaug and Peter Bajcsy National Center for Supercomputing pplications 605 East Springfield venue, Campaign,

More information

Tilings of rectangles with T-tetrominoes

Tilings of rectangles with T-tetrominoes Tilings of rectangles wit T-tetrominoes Micael Korn and Igor Pak Department of Matematics Massacusetts Institute of Tecnology Cambridge, MA, 2139 mikekorn@mit.edu, pak@mat.mit.edu August 26, 23 Abstract

More information

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Fall

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Fall Has-Based Indexes Capter 11 Comp 521 Files and Databases Fall 2012 1 Introduction Hasing maps a searc key directly to te pid of te containing page/page-overflow cain Doesn t require intermediate page fetces

More information

Real-Time Wireless Routing for Industrial Internet of Things

Real-Time Wireless Routing for Industrial Internet of Things Real-Time Wireless Routing for Industrial Internet of Tings Cengjie Wu, Dolvara Gunatilaka, Mo Sa, Cenyang Lu Cyber-Pysical Systems Laboratory, Wasington University in St. Louis Department of Computer

More information

Classification of Osteoporosis using Fractal Texture Features

Classification of Osteoporosis using Fractal Texture Features Classification of Osteoporosis using Fractal Texture Features V.Srikant, C.Dines Kumar and A.Tobin Department of Electronics and Communication Engineering Panimalar Engineering College Cennai, Tamil Nadu,

More information

Section 2.3: Calculating Limits using the Limit Laws

Section 2.3: Calculating Limits using the Limit Laws Section 2.3: Calculating Limits using te Limit Laws In previous sections, we used graps and numerics to approimate te value of a it if it eists. Te problem wit tis owever is tat it does not always give

More information

A Novel QC-LDPC Code with Flexible Construction and Low Error Floor

A Novel QC-LDPC Code with Flexible Construction and Low Error Floor A Novel QC-LDPC Code wit Flexile Construction and Low Error Floor Hanxin WANG,2, Saoping CHEN,2,CuitaoZHU,2 and Kaiyou SU Department of Electronics and Information Engineering, Sout-Central University

More information

Optimal In-Network Packet Aggregation Policy for Maximum Information Freshness

Optimal In-Network Packet Aggregation Policy for Maximum Information Freshness 1 Optimal In-etwork Packet Aggregation Policy for Maimum Information Fresness Alper Sinan Akyurek, Tajana Simunic Rosing Electrical and Computer Engineering, University of California, San Diego aakyurek@ucsd.edu,

More information

CHAPTER 7: TRANSCENDENTAL FUNCTIONS

CHAPTER 7: TRANSCENDENTAL FUNCTIONS 7.0 Introduction and One to one Functions Contemporary Calculus 1 CHAPTER 7: TRANSCENDENTAL FUNCTIONS Introduction In te previous capters we saw ow to calculate and use te derivatives and integrals of

More information

Non-Interferometric Testing

Non-Interferometric Testing NonInterferometric Testing.nb Optics 513 - James C. Wyant 1 Non-Interferometric Testing Introduction In tese notes four non-interferometric tests are described: (1) te Sack-Hartmann test, (2) te Foucault

More information

Minimizing Memory Access By Improving Register Usage Through High-level Transformations

Minimizing Memory Access By Improving Register Usage Through High-level Transformations Minimizing Memory Access By Improving Register Usage Troug Hig-level Transformations San Li Scool of Computer Engineering anyang Tecnological University anyang Avenue, SIGAPORE 639798 Email: p144102711@ntu.edu.sg

More information

Soft sensor modelling by time difference, recursive partial least squares and adaptive model updating

Soft sensor modelling by time difference, recursive partial least squares and adaptive model updating Soft sensor modelling by time difference, recursive partial least squares adaptive model updating Y Fu 1, 2, W Yang 2, O Xu 1, L Zou 3, J Wang 4 1 Zijiang College, Zejiang University of ecnology, Hangzou

More information

SLOTTED-RING LOCAL AREA NETWORKS WITH MULTIPLE PRIORITY STATIONS. Hewlett-Packard Company East Mission Avenue. Bogazici University

SLOTTED-RING LOCAL AREA NETWORKS WITH MULTIPLE PRIORITY STATIONS. Hewlett-Packard Company East Mission Avenue. Bogazici University SLOTTED-RING LOCAL AREA NETWORKS WITH MULTIPLE PRIORITY STATIONS Sanuj V. Sarin 1, Hakan Delic 2 and Jung H. Kim 3 1 Hewlett-Packard Company 24001 East Mission Avenue Spokane, Wasington 99109, USA 2 Signal

More information

16th European Signal Processing Conference (EUSIPCO 2008), Lausanne, Switzerland, August 25-29, 2008, copyright by EURASIP

16th European Signal Processing Conference (EUSIPCO 2008), Lausanne, Switzerland, August 25-29, 2008, copyright by EURASIP 16t European Signal Processing Conference (EUSIPCO 008), Lausanne, Switzerland, August 5-9, 008, copyrigt by EURASIP ADAPTIVE WINDOW FOR LOCAL POLYNOMIAL REGRESSION FROM NOISY NONUNIFORM SAMPLES A. Sreenivasa

More information

Proceedings of the 8th WSEAS International Conference on Neural Networks, Vancouver, British Columbia, Canada, June 19-21,

Proceedings of the 8th WSEAS International Conference on Neural Networks, Vancouver, British Columbia, Canada, June 19-21, Proceedings of te 8t WSEAS International Conference on Neural Networks, Vancouver, Britis Columbia, Canada, June 9-2, 2007 3 Neural Network Structures wit Constant Weigts to Implement Dis-Jointly Removed

More information

Interference and Diffraction of Light

Interference and Diffraction of Light Interference and Diffraction of Ligt References: [1] A.P. Frenc: Vibrations and Waves, Norton Publ. 1971, Capter 8, p. 280-297 [2] PASCO Interference and Diffraction EX-9918 guide (written by Ann Hanks)

More information

CRASHWORTHINESS ASSESSMENT IN AIRCRAFT DITCHING INCIDENTS

CRASHWORTHINESS ASSESSMENT IN AIRCRAFT DITCHING INCIDENTS 27 TH INTERNATIONAL CONGRESS OF THE AERONAUTICAL SCIENCES CRASHWORTHINESS ASSESSMENT IN AIRCRAFT DITCHING INCIDENTS C. Candra*, T. Y. Wong* and J. Bayandor** * Te Sir Lawrence Wackett Aerospace Centre

More information

RECONSTRUCTING OF A GIVEN PIXEL S THREE- DIMENSIONAL COORDINATES GIVEN BY A PERSPECTIVE DIGITAL AERIAL PHOTOS BY APPLYING DIGITAL TERRAIN MODEL

RECONSTRUCTING OF A GIVEN PIXEL S THREE- DIMENSIONAL COORDINATES GIVEN BY A PERSPECTIVE DIGITAL AERIAL PHOTOS BY APPLYING DIGITAL TERRAIN MODEL IV. Évfolyam 3. szám - 2009. szeptember Horvát Zoltán orvat.zoltan@zmne.u REONSTRUTING OF GIVEN PIXEL S THREE- DIMENSIONL OORDINTES GIVEN Y PERSPETIVE DIGITL ERIL PHOTOS Y PPLYING DIGITL TERRIN MODEL bsztrakt/bstract

More information

Feature-Based Steganalysis for JPEG Images and its Implications for Future Design of Steganographic Schemes

Feature-Based Steganalysis for JPEG Images and its Implications for Future Design of Steganographic Schemes Feature-Based Steganalysis for JPEG Images and its Implications for Future Design of Steganograpic Scemes Jessica Fridric Dept. of Electrical Engineering, SUNY Bingamton, Bingamton, NY 3902-6000, USA fridric@bingamton.edu

More information

Asynchronous Power Flow on Graphic Processing Units

Asynchronous Power Flow on Graphic Processing Units 1 Asyncronous Power Flow on Grapic Processing Units Manuel Marin, Student Member, IEEE, David Defour, and Federico Milano, Senior Member, IEEE Abstract Asyncronous iterations can be used to implement fixed-point

More information

Nonprehensile Manipulation for Orienting Parts in the Plane

Nonprehensile Manipulation for Orienting Parts in the Plane Nonpreensile Manipulation for Orienting Parts in te Plane Nina B. Zumel Robotics Institute Carnegie Mellon University zumel@ri.cmu.edu Micael A. Erdmann Robotics Institute and Scool of Computer Science

More information

Materials: Whiteboard, TI-Nspire classroom set, quadratic tangents program, and a computer projector.

Materials: Whiteboard, TI-Nspire classroom set, quadratic tangents program, and a computer projector. Adam Clinc Lesson: Deriving te Derivative Grade Level: 12 t grade, Calculus I class Materials: Witeboard, TI-Nspire classroom set, quadratic tangents program, and a computer projector. Goals/Objectives:

More information

Solutions Manual for Fundamentals of Fluid Mechanics 7th edition by Munson Rothmayer Okiishi and Huebsch

Solutions Manual for Fundamentals of Fluid Mechanics 7th edition by Munson Rothmayer Okiishi and Huebsch Solutions Manual for Fundamentals of Fluid Mecanics 7t edition by Munson Rotmayer Okiisi and Huebsc Link full download : ttps://digitalcontentmarket.org/download/solutions-manual-forfundamentals-of-fluid-mecanics-7t-edition-by-munson-rotmayer-okiisi-and-uebsc/

More information

An Effective Sensor Deployment Strategy by Linear Density Control in Wireless Sensor Networks Chiming Huang and Rei-Heng Cheng

An Effective Sensor Deployment Strategy by Linear Density Control in Wireless Sensor Networks Chiming Huang and Rei-Heng Cheng An ffective Sensor Deployment Strategy by Linear Density Control in Wireless Sensor Networks Ciming Huang and ei-heng Ceng 5 De c e mbe r0 International Journal of Advanced Information Tecnologies (IJAIT),

More information

Traffic Pattern-based Adaptive Routing for Intra-group Communication in Dragonfly Networks

Traffic Pattern-based Adaptive Routing for Intra-group Communication in Dragonfly Networks Traffic Pattern-based Adaptive Routing for Intra-group Communication in Dragonfly Networks Peyman Faizian, Md Safayat Raman, Md Atiqul Molla, Xin Yuan Department of Computer Science Florida State University

More information

George Xylomenos and George C. Polyzos. with existing protocols and their eæciency in terms of

George Xylomenos and George C. Polyzos. with existing protocols and their eæciency in terms of IP MULTICASTING FOR WIRELESS MOBILE OSTS George Xylomenos and George C. Polyzos fxgeorge,polyzosg@cs.ucsd.edu Computer Systems Laboratory Department of Computer Science and Engineering University of California,

More information

Search-aware Conditions for Probably Approximately Correct Heuristic Search

Search-aware Conditions for Probably Approximately Correct Heuristic Search Searc-aware Conditions for Probably Approximately Correct Heuristic Searc Roni Stern Ariel Felner Information Systems Engineering Ben Gurion University Beer-Seva, Israel 85104 roni.stern@gmail.com, felner@bgu.ac.il

More information

HASH ALGORITHMS: A DESIGN FOR PARALLEL CALCULATIONS

HASH ALGORITHMS: A DESIGN FOR PARALLEL CALCULATIONS HASH ALGORITHMS: A DESIGN FOR PARALLEL CALCULATIONS N.G.Bardis Researc Associate Hellenic Ministry of te Interior, Public Administration and Decentralization 8, Dragatsaniou str., Klatmonos S. 0559, Greece

More information

Some Handwritten Signature Parameters in Biometric Recognition Process

Some Handwritten Signature Parameters in Biometric Recognition Process Some Handwritten Signature Parameters in Biometric Recognition Process Piotr Porwik Institute of Informatics, Silesian Uniersity, Bdziska 39, 41- Sosnowiec, Poland porwik@us.edu.pl Tomasz Para Institute

More information

Alternating Direction Implicit Methods for FDTD Using the Dey-Mittra Embedded Boundary Method

Alternating Direction Implicit Methods for FDTD Using the Dey-Mittra Embedded Boundary Method Te Open Plasma Pysics Journal, 2010, 3, 29-35 29 Open Access Alternating Direction Implicit Metods for FDTD Using te Dey-Mittra Embedded Boundary Metod T.M. Austin *, J.R. Cary, D.N. Smite C. Nieter Tec-X

More information

Mean Waiting Time Analysis in Finite Storage Queues for Wireless Cellular Networks

Mean Waiting Time Analysis in Finite Storage Queues for Wireless Cellular Networks Mean Waiting Time Analysis in Finite Storage ueues for Wireless ellular Networks J. YLARINOS, S. LOUVROS, K. IOANNOU, A. IOANNOU 3 A.GARMIS 2 and S.KOTSOOULOS Wireless Telecommunication Laboratory, Department

More information

An Anchor Chain Scheme for IP Mobility Management

An Anchor Chain Scheme for IP Mobility Management An Ancor Cain Sceme for IP Mobility Management Yigal Bejerano and Israel Cidon Department of Electrical Engineering Tecnion - Israel Institute of Tecnology Haifa 32000, Israel E-mail: bej@tx.tecnion.ac.il.

More information

Network Coding to Enhance Standard Routing Protocols in Wireless Mesh Networks

Network Coding to Enhance Standard Routing Protocols in Wireless Mesh Networks Downloaded from vbn.aau.dk on: April 7, 09 Aalborg Universitet etwork Coding to Enance Standard Routing Protocols in Wireless Mes etworks Palevani, Peyman; Roetter, Daniel Enrique Lucani; Fitzek, Frank;

More information

A UPnP-based Decentralized Service Discovery Improved Algorithm

A UPnP-based Decentralized Service Discovery Improved Algorithm Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol.1, No.1, Marc 2013, pp. 21~26 ISSN: 2089-3272 21 A UPnP-based Decentralized Service Discovery Improved Algoritm Yu Si-cai*, Wu Yan-zi,

More information

Implementation of Integral based Digital Curvature Estimators in DGtal

Implementation of Integral based Digital Curvature Estimators in DGtal Implementation of Integral based Digital Curvature Estimators in DGtal David Coeurjolly 1, Jacques-Olivier Lacaud 2, Jérémy Levallois 1,2 1 Université de Lyon, CNRS INSA-Lyon, LIRIS, UMR5205, F-69621,

More information

Analytical CHEMISTRY

Analytical CHEMISTRY ISSN : 974-749 Grap kernels and applications in protein classification Jiang Qiangrong*, Xiong Zikang, Zai Can Department of Computer Science, Beijing University of Tecnology, Beijing, (CHINA) E-mail:

More information

SlidesGen: Automatic Generation of Presentation Slides for a Technical Paper Using Summarization

SlidesGen: Automatic Generation of Presentation Slides for a Technical Paper Using Summarization Proceedings of te Twenty-Second International FLAIRS Conference (2009) SlidesGen: Automatic Generation of Presentation Slides for a Tecnical Paper Using Summarization M. Sravanti, C. Ravindranat Cowdary

More information

ICES REPORT Isogeometric Analysis of Boundary Integral Equations

ICES REPORT Isogeometric Analysis of Boundary Integral Equations ICES REPORT 5-2 April 205 Isogeometric Analysis of Boundary Integral Equations by Mattias Taus, Gregory J. Rodin and Tomas J. R. Huges Te Institute for Computational Engineering and Sciences Te University

More information

PYRAMID FILTERS BASED ON BILINEAR INTERPOLATION

PYRAMID FILTERS BASED ON BILINEAR INTERPOLATION PYRAMID FILTERS BASED ON BILINEAR INTERPOLATION Martin Kraus Computer Grapics and Visualization Group, Tecnisce Universität Müncen, Germany krausma@in.tum.de Magnus Strengert Visualization and Interactive

More information

Fault Localization Using Tarantula

Fault Localization Using Tarantula Class 20 Fault localization (cont d) Test-data generation Exam review: Nov 3, after class to :30 Responsible for all material up troug Nov 3 (troug test-data generation) Send questions beforeand so all

More information

Cooperation in Wireless Ad Hoc Networks

Cooperation in Wireless Ad Hoc Networks Cooperation in Wireless Ad Hoc Networks Vikram Srinivasan, Pavan Nuggealli, Carla F. Ciasserini, Rames R. Rao Department of Electrical and Computer Engineering, University of California at San Diego email:

More information

Protecting Storage Location Privacy in Sensor Networks

Protecting Storage Location Privacy in Sensor Networks Protecting Storage Location Privacy in Sensor Networks Jianming Zou, Wenseng Zang, and Daji Qiao Iowa State University Ames, IA 511 Email: {jmzou,wzang,daji}@iastate.edu ASTAT Numerous scemes ave been

More information

MAPI Computer Vision

MAPI Computer Vision MAPI Computer Vision Multiple View Geometry In tis module we intend to present several tecniques in te domain of te 3D vision Manuel Joao University of Mino Dep Industrial Electronics - Applications -

More information

Laser Radar based Vehicle Localization in GPS Signal Blocked Areas

Laser Radar based Vehicle Localization in GPS Signal Blocked Areas International Journal of Computational Intelligence Systems, Vol. 4, No. 6 (December, 20), 00-09 Laser Radar based Veicle Localization in GPS Signal Bloced Areas Ming Yang Department of Automation, Sangai

More information

Grid Adaptation for Functional Outputs: Application to Two-Dimensional Inviscid Flows

Grid Adaptation for Functional Outputs: Application to Two-Dimensional Inviscid Flows Journal of Computational Pysics 176, 40 69 (2002) doi:10.1006/jcp.2001.6967, available online at ttp://www.idealibrary.com on Grid Adaptation for Functional Outputs: Application to Two-Dimensional Inviscid

More information

UUV DEPTH MEASUREMENT USING CAMERA IMAGES

UUV DEPTH MEASUREMENT USING CAMERA IMAGES ABCM Symposium Series in Mecatronics - Vol. 3 - pp.292-299 Copyrigt c 2008 by ABCM UUV DEPTH MEASUREMENT USING CAMERA IMAGES Rogerio Yugo Takimoto Graduate Scool of Engineering Yokoama National University

More information

A Bidirectional Subsethood Based Similarity Measure for Fuzzy Sets

A Bidirectional Subsethood Based Similarity Measure for Fuzzy Sets A Bidirectional Subsetood Based Similarity Measure for Fuzzy Sets Saily Kabir Cristian Wagner Timoty C. Havens and Derek T. Anderson Intelligent Modelling and Analysis (IMA) Group and Lab for Uncertainty

More information

Software Fault Prediction using Machine Learning Algorithm Pooja Garg 1 Mr. Bhushan Dua 2

Software Fault Prediction using Machine Learning Algorithm Pooja Garg 1 Mr. Bhushan Dua 2 IJSRD - International Journal for Scientific Researc & Development Vol. 3, Issue 04, 2015 ISSN (online): 2321-0613 Software Fault Prediction using Macine Learning Algoritm Pooja Garg 1 Mr. Busan Dua 2

More information

The navigability variable is binary either a cell is navigable or not. Thus, we can invert the entire reasoning by substituting x i for x i : (4)

The navigability variable is binary either a cell is navigable or not. Thus, we can invert the entire reasoning by substituting x i for x i : (4) A Multi-Resolution Pyramid for Outdoor Robot Terrain Perception Micael Montemerlo and Sebastian Trun AI Lab, Stanford University 353 Serra Mall Stanford, CA 94305-9010 {mmde,trun}@stanford.edu Abstract

More information

Efficient Content-Based Indexing of Large Image Databases

Efficient Content-Based Indexing of Large Image Databases Efficient Content-Based Indexing of Large Image Databases ESSAM A. EL-KWAE University of Nort Carolina at Carlotte and MANSUR R. KABUKA University of Miami Large image databases ave emerged in various

More information

Packet Switching Networks. Jonathan S. Turner. Computer and Communications Research Center. the result of user connections that pass through the

Packet Switching Networks. Jonathan S. Turner. Computer and Communications Research Center. the result of user connections that pass through the Fluid Flow Loading Analysis of Packet Switcing Networks Jonatan S. Turner Computer and Communications Researc Center Wasington University, St. Louis, MO 63130 Abstract Recent researc in switcing as concentrated

More information